Add healthcheck in Dockerfile

[mik-laj]

Part of: Koenkk/zigbee2mqtt#28893

This adds a basic health check.
A more complete check (including whether the HTTP server is connected to MQTT) would be useful, but that’ll require server-side changes — so it’s better left for a follow-up.
Even so, this should already improve web server stability, since the Supervisor won’t route traffic to the app until it can accept HTTP requests.

[Copilot]

Pull Request Overview

This PR adds a basic HTTP health check to the Dockerfile to improve web server stability by ensuring the container can accept HTTP requests before routing traffic to it.

• Adds validation to confirm curl is available in the final Docker image
• Implements a health check that tests HTTP connectivity on port 8099
• Configures health check with 1-minute intervals and 3-second timeout

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Corrected capitalization of 'CURL' to 'curl'.

Suggested change

	# Validation to confirm that CURL is installed in the final image.
	# Validation to confirm that curl is installed in the final image.

[Copilot]

Pull Request Overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 2 comments.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Koenkk]

Thanks!

[Nerivec]

Has this been tested with onboarding?
See Koenkk/zigbee2mqtt#29128 (comment)

[mik-laj]

Unfortunately not. I was just testing if the command ran correctly and I used docker exec for that.

[Koenkk]

I just checked and it works, also made some improvement to make the healthcheck pass faster during startup (1ebfd6f) (otherwise it takes 1 minute before the frontend is accessible)

[Nerivec]

otherwise it takes 1 minute before the frontend is accessible

The healthcheck is preventing access until it passes??

---

The race condition (healthcheck happening exactly between the switch from onboarding to Controller -frontend started-) would be nearly impossible to trigger manually. With a good instance, we're talking just a few seconds, but enough that it will undoubtedly trigger many times in real-world scenario.

[mik-laj]

The healthcheck is preventing access until it passes??

Yes. In the case of HA Adoonu. Supervisor does not forward traffic to the add-on, but displays a message that the application is not available.

[Koenkk]

With a good instance, we're talking just a few seconds, but enough that it will undoubtedly trigger many times in real-world scenario.

It might trigger in real world indeed, but this will only prevent users from accessing the frontend for 1 minute after onboarding (something a user only does typically once). So I think that is acceptable.